Macromolecule mediated transport and retention of · 2010. 4. 13. · Macromolecule mediated...
Transcript of Macromolecule mediated transport and retention of · 2010. 4. 13. · Macromolecule mediated...
w a t e r r e s e a r c h 4 4 ( 2 0 1 0 ) 1 0 8 2 – 1 0 9 3
Avai lab le at www.sc iencedi rect .com
journa l homepage : www.e lsev i er . com/ loca te /wat res
Macromolecule mediated transport and retention ofEscherichia coli O157:H7 in saturated porous media
Hyunjung N. Kim a, Sharon L. Walker a, Scott A. Bradford b,*a Department of Chemical and Environmental Engineering, University of California, Riverside, CA 92521, USAb USDA, ARS, US Salinity Laboratory, Riverside, CA 92507, USA
a r t i c l e i n f o
Article history:
Received 28 January 2009
Received in revised form
1 September 2009
Accepted 9 September 2009
Available online 12 September 2009
Keywords:
Pathogen
Transport
Escherichia coli O157:H7
Extracellular macromolecules
Polymer bridging
Steric stabilization
* Corresponding author. Tel.: þ1 951 369 485E-mail address: [email protected]
0043-1354/$ – see front matter Published bydoi:10.1016/j.watres.2009.09.027
a b s t r a c t
The role of extracellular macromolecules on Escherichia coli O157:H7 transport and reten-
tion was investigated in saturated porous media. To compare the relative transport and
retention of E. coli cells that are macromolecule rich and deficient, macromolecules were
partially cleaved using a proteolytic enzyme. Characterization of bacterial cell surfaces, cell
aggregation, and experiments in a packed sand column were conducted over a range of
ionic strength (IS). The results showed that macromolecule-related interactions contribute
to retention of E. coli O157:H7 and are strongly linked to solution IS. Under low IS conditions
(IS � 0.1 mM), partial removal of the macromolecules resulted in a more negative elec-
trophoretic mobility of cells and created more unfavorable conditions for cell–quartz and
cell–cell interactions as suggested by Derjaguin-Landau-Verwey-Overbeek (DLVO) inter-
action energy profiles and cell aggregation kinetics. Consequently, less retention was
observed for enzyme treated cells in the corresponding column experiments. In addition,
a time-dependent deposition process (i.e., ripening) was observed for untreated cells, but
not for treated cells, supporting the fact that the macromolecules enhanced cell–cell
interactions. Additional column experiments for untreated cells under favorable condi-
tions (IS � 1 mM) showed that a significant amount of the cells were reversibly retained in
the column, which contradicts predictions of DLVO theory. Furthermore, a non-monotonic
cell retention profile was observed under favorable attachment conditions. These obser-
vations indicated that the presence of macromolecules hindered irreversible interactions
between the cells and the quartz surface.
Published by Elsevier Ltd.
1. Introduction per year are estimated to occur in the United States due to E.
Escherichia coli O157:H7, a predominant serotype of enter-
ohemorrhagic E. coli, is a Gram-negative pathogenic bacte-
rium. This bacterium can cause severe gastrointestinal
disease in humans, such as bloody diarrhea and hemolytic
uremic syndrome, which may result in acute renal failure in
children (Boyce et al., 1995; Dean-Nystrom et al., 2003; Kaper
and Karmali, 2008). More than 70,000 illnesses and 60 deaths
7; fax: þ1 951 342 4964.ov (S.A. Bradford).Elsevier Ltd.
coli O157:H7 infection (Mead et al., 1999).
Pathogens can reach drinking water supplies via a number
of pathways; e.g., surface runoff, soil infiltration, and
groundwater recharge. The transmission of waterborne
diseases by pathogen-contaminated groundwater is a growing
concern. An accurate understanding of the transport and fate
of pathogens in subsurface and groundwater environments is
needed to protect water resources. Much effort has been
w a t e r r e s e a r c h 4 4 ( 2 0 1 0 ) 1 0 8 2 – 1 0 9 3 1083
devoted to understand fundamental mechanisms of fate and
transport of colloids and microorganisms in porous media
(Schijven and Hassanizadeh, 2000; Ginn et al., 2002; Jin and
Flury, 2002). The transport and deposition of non-living
colloids has frequently been shown to be influenced by
physicochemical phenomena between the interacting
surfaces (Bradford et al., 2007; Kuznar and Elimelech, 2007;
Shen et al., 2007). Unlike the non-living colloids, bacteria are
living organisms that exhibit distinct characteristics. For
example, bacteria may be rod-shaped and physiologically
respond to environmental changes (Frank, 2001; Seltmann
and Holst, 2002). Most bacteria possess various types of
surface-bound macromolecules such as proteins, lipopoly-
saccharides, fimbriae, flagella (Frank, 2001; Seltmann and
Holst, 2002). In addition, unbound macromolecules (e.g.,
extracellular polymeric substances) are frequently found
outside cell surface (Wingender et al., 1999). These macro-
molecules have very complex structure, and the type,
composition, and amount of macromolecules can vary with
other factors, such as bacterial species, nutrient conditions,
growth stage, etc. (Law, 2000; Walker et al., 2005; Chen and
Walker, 2007).
A growing body of literature indicates that surface/extra-
cellular macromolecules can play a significant role in the
interaction of microbes and abiotic surfaces (Kuznar and Eli-
melech, 2005; Liu et al., 2007; Gargiulo et al., 2007). The pres-
ence of macromolecules on/outside the surface of microbes
may cause steric forces (Israelachvili, 1992; van Oss, 1994).
Several studies reported that surface/extracellular macro-
molecules can either enhance (Jucker et al., 1998; Rijnaarts
et al., 1999; Gargiulo et al., 2007) or hinder (Jucker et al., 1998;
Kuznar and Elimelech, 2005, 2006) the adhesion of microbes to
abiotic surfaces. Studies have also reported that the presence
of bacterial polymeric layers influence cell–cell interactions
due to physicochemical (Voloshin and Kaprelyants, 2004) or
biological (Frank, 2001; Voloshin and Kaprelyants, 2004)
mechanisms. Furthermore, cell–cell interactions have also
been proposed to be an important mechanism in bacteria
retention in porous media (Bradford et al., 2006b; Kim et al.,
2009a), in cell adhesion to biotic/abiotic surfaces (Frank, 2001),
and in bacterial colonization/biofilm formation (Frank, 2001;
Voloshin and Kaprelyants, 2004). Although much research has
been devoted to investigate the macromolecule-mediated
interactions (i.e., cell–cell and cell–surface interactions), the
fundamental interaction mechanisms are not well
understood because the type and extent of the macromole-
cule-mediated interactions depend on the cell type and the
environmental conditions (e.g., pH, IS, temperature). Further
study is required to better elucidate these interactions.
E. coli O157:H7 has been reported to produce extracellular
polymeric materials, which may protect the cells from harsh
environments and initiate cell colonization/biofilm formation
(Ryu and Beuchat, 2005; Oh et al., 2007). Previously, we observed
that the E. coli strain selected for this study produced extra-
cellular materials (Kim et al., 2009b). To directly investigate the
influence of macromolecules on the phenotypic properties
(e.g., electrophoretic mobility (EPM), hydrophobicity, acid–base
property, etc.) of E. coli O157:H7 cells some cells were treated
with a proteolytic enzyme (i.e., proteinase K) to cleave exposed
extracellular macromolecules. The study showed that the
presence of the extracellular macromolecules altered the cell
surface characteristics and eventually influenced the adhesion
behavior of the cells in a batch system at pH ¼ 5.8 (Kim et al.,
2009b). Another recent study showed that E. coli O157:H7
retention in a packed sand column was inversely proportional
to IS under high pH conditions (IS¼ 1–100 mM and pH¼ 8.4–9.2)
(Kim et al., 2009a). This behavior was not explained by classic
DLVO theory (Derjaguin and Landau, 1941; Verwey and Over-
beek, 1948). Instead, the trend was attributed to pH-associated
steric stabilization (Kim et al., 2009a), indirectly suggesting the
presence of potential macromolecule-related interactions.
Although Kim et al. (2009a,b) provide useful insight about
the role of extracellular macromolecules on E. coli O157:H7
interactions and transport, gaps in knowledge still remain
that are addressed in the current work. Specifically, we
investigate the potential influence of E. coli O157:H7 extracel-
lular macromolecules on: (a) attractive steric interactions
under low IS conditions (i.e., polymer bridging); and (b)
repulsive steric interactions under low pH conditions (effects
are much more subtle at pH ¼ 5.8 than at higher pH condi-
tions). The implications of these steric interactions on cell
transport and aggregation is subsequently quantified. For this
purpose, a proteolytic enzyme (i.e., proteinase K) was
employed to partially cleave the macromolecules on the E. coli
O157:H7 cell surface, and the transport and retention behavior
of the proteinase K treated E. coli O157:H7 cells was examined
and compared with the untreated cells in a packed sand
column over the IS range of 0.01–100 mM KCl (pH 5.8). In
addition, aggregation tests were conducted with the untreated
and proteinase K treated E. coli O157:H7 cells as a function of IS
to examine the relationship between cell–cell interaction and
cell retention behavior. A similar study has been previously
conducted by Kuznar and Elimelech (2006) using Cryptospo-
ridium parvum oocysts. They compared the attachment effi-
ciency between untreated and proteinase K treated oocysts to
quartz surface in a radial stagnation point flow (RSPF) system.
However, the RSPF system only captured irreversible inter-
actions and they focused on repulsive steric interactions. In
addition, we are unaware of any other studies that have
applied this enzymatic technique for pathogenic E. coli
O157:H7 in a column system.
2. Materials and methods
2.1. Bacterial growth and preparation
E. coli O157:H7/pGFP strain 72 was obtained from Dr. Pina
Fratamico (USDA-ARS-ERRC, Wyndmoor, PA) for this study.
The cells were precultured in Tryptic Soy Broth (TSB, Becton
Dickinson, Sparks, MD) in the presence of 0.1 g/L ampicillin
(Sigma–Aldrich, St. Louis, MO) at 200 rpm and 37 �C in an
incubator (Model 4639, Barnstead/Labline, Melrose Park, IL).
The precultured cells were transferred onto Tryptic Soy Agar
(TSA, Becton Dickinson) in Petri dishes with 0.1 g/L ampicillin
(Sigma–Aldrich), and grown for 18 h (stationary phase) at 37�C. In order to collect the cultured cells, sterile de-ionized (DI)
water was added into TSA and the cells were gently scraped
using a sterile cell spreader (Fisher Scientific, Fair Lawn, NJ).
The collected cells were harvested by centrifugation (Fisher
w a t e r r e s e a r c h 4 4 ( 2 0 1 0 ) 1 0 8 2 – 1 0 9 31084
accuSpin* 3R Centrifuge, Fisher Scientific) for 15 min at 3700 g
(Swing Bucket Rotor 7500, Fisher Scientific). The cells were
rinsed two additional times using a 10 mM potassium chloride
(KCl, Fisher Scientific) solution to remove any remaining
growth media, and then resuspended in 3 mL of the same
electrolyte solution chemistry of the subsequent experiment.
This cell suspension was used to determine cell concentration
in a counting chamber (Marienfeld Laboratory Glassware,
Germany).
To investigate the role of extracellular macromolecules on
E. coli O157:H7 transport and retention in a packed sand
column, a proteolytic enzyme (proteinase K, Sigma–Aldrich)
was employed to partially cleave the extracellular macro-
molecules from the bacteria. A detailed description of this
protocol is available in the literature (Kuznar and Elimelech,
2005; Kim et al., 2009b). Briefly, the untreated cells harvested
by the above protocol were suspended into a background
solution, which consists of 0.1 mg/mL proteinase K (Sigma–
Aldrich), 5 mM EDTA, 0.01 g/L sodium dodecyl sulfate, and 10
mM Tris–HCl (pH 8.0). The final cell concentration was
approximately 5 � 108 cells/mL. The cell suspension was
incubated at 37 �C and 260 rpm for 3 h to digest macromole-
cules on/outside the cell surface. The proteinase K treated
cells were then rinsed twice with DI water by a centrifugation
at 13400 g for 2 min, followed by one more rinsing with DI
water at 3700 g for 10 min. All chemicals for the digestion
process were reagent grade (Sigma–Aldrich).
2.2. Solution chemistry and bacterial characterization
KCl was selected as the background electrolyte for the E. coli
O157:H7 characterization and transport experiments, and the
IS ranged from 0.01 to 100 mM. The pH for all experiments was
maintained at approximately 5.8.
Cell viability, size, and shape were examined as a function
of IS for untreated and proteinase K treated E. coli O157:H7
cells. Viability tests were performed based on the Live/Dead
BacLight (L-7012, Molecular Probes, Eugene, OR) method
(Boulos et al., 1999) with an inverted fluorescent microscope
(IX70, Olympus, Japan) equipped with a red/green fluores-
cence filter set (Chroma Technology Corp., Brattleboro, VT).
Bacterial cell size and shape were also determined from
images taken with the microscope in phase contrast mode.
Volume-based equivalent radii and aspect ratios of the cells
were determined from the measured lengths and widths of
Table 1 – Electrophoretic mobility of E. coli O157:H7 cells as wel(n-dodecane) before and after proteinase K treatment.a
Ionic strength (mM) Electrophoretic mobility(mm V s�1 cm�1)b
Untreated Proteinase
0.01 �1.234 � 0.135 �3.086 �0.1 �0.644 � 0.080 �2.176 �1 �0.073 � 0.147 �0.705 �100 �0.112 � 0.742 �0.053 �
a The experiments were conducted at pH 5.8 and at room temperature (
b Data from Kim et al. (2009b). ND Not determined.
the cells (n � 50). The percentage of viable cells were deter-
mined to be more than 90% regardless of IS and enzyme
treatment. The average effective cell diameter was approxi-
mately 0.71 mm (minimum and maximum value of 0.66 and
0.78 mm, respectively), and the average aspect ratio also
ranged from 2.9 � 1.6 (minimum) to 3.7 � 2.2 (maximum). The
detailed results for cell viability, size, and shape measure-
ments can be found in Kim et al. (2009b) and indicate that
proteinase K did not damage the cell wall during the macro-
molecule treatment and there was no significant impact of
osmotic pressure on E. coli O157:H7 cells at low and high IS.
EPM of the untreated and proteinase K treated E. coli
O157:H7 cells were evaluated as a function of IS (0.01–100 mM).
Detailed information regarding the procedures can be found
in Kim et al. (2009b). In brief, EPM measurements were con-
ducted at 25 �C using a ZetaPALS analyzer (Brookhaven
Instruments Corporation, Holtsville, NY). The pH of the solu-
tion was unadjusted. Three different samples were prepared
for each solution IS condition and average values of ten runs
were obtained for one sample.
Hydrophobicity tests were conducted for the untreated and
proteinase K treated E. coli O157:H7 cells at 0.01 and 0.1 mM (pH
5.8) based on the microbial adhesion to hydrocarbons (MATH)
test (Pembrey et al., 1999). One mL of n-dodecane (laboratory
grade, Fisher Scientific) was added to 4 mL of a cell suspension,
and the suspension was mixed for 2 min. The mixture was then
left for at least 15 min at room temperature to allow the two
liquids to separate. The percentage of the cells partitioned to the
hydrocarbon phase was calculated after the optical density of
the cell suspension in the water phase was determined at 546
nm (BioSpec-mini, Shimadzu Corp.). All experiments were
conducted at least in triplicate. The results are presented in
Table 1. Overall, the proteinase K treated cells were determined
to be slightly more hydrophilic than the untreated ones.
Statistical differences between mean values were analyzed
using a student t-test. When P < 0.05, the differences are
considered to be statistically significant.
2.3. Cell aggregation tests
In order to evaluate the initial aggregation kinetics of
untreated and proteinase K treated E. coli O157:H7 cells at
different IS, the cells were harvested and washed according to
the procedure in Section 2.1. For the untreated cells, the
washed pellets were resuspended in the select KCl solution
l as the percentage of the cells partitioned into hydrocarbon
Hydrophobicity (%)
K treated Untreated Proteinase K treated
0.265 15.5 � 2.8 11.3 � 1.8
0.071 14.5 � 1.3 11.6 � 1.7
0.187 ND ND
0.557 ND ND
22–25 �C).
w a t e r r e s e a r c h 4 4 ( 2 0 1 0 ) 1 0 8 2 – 1 0 9 3 1085
(IS ¼ 0.01, 0.1, 1, and 100 mM) to an optical density
(i.e., absorbance value) of 0.6 (dimensionless unit) at a wave-
length of 600 nm, which corresponds to the cell concentration
of 108 cells/mL. The cell suspension (0.4 mL) was transferred
into a quartz cuvette (Fisher Scientific) and the upper part of
the sample was measured spectroscopically (SP-890, Barn-
stead International, Dubuque, IA) at the wavelength of 600 nm
for 2 h. The aggregation tests for the treated cells were also
carried out in the presence of 0.01 and 0.1 mM KCl after the
enzyme treatment.
2.4. Porous medium preparation
Ultra-pure quartz sand (Iota� quartz, Unimin Corp., NC),
which has the average diameter (d50) of 275 mm, was chosen
for column experiments. To remove any metal and organic
impurities, the sand was thoroughly cleaned using the
method of Litton and Olson (1993). Prior to wet-packing the
column, the cleaned sand was re-hydrated by boiling in DI
water for 1 h. The total (or effective) porosity was determined
to be ca. 0.46 by a gravimetric method. The zeta potential of
the quartz sand was measured at IS of 0.01 and 0.1 mM (KCl)
using an Electro Kinetic Analyzer (Anton Paar GmbH, Graz,
Austria) equipped with a cylindrical cell to calculate DLVO
interaction energy between the cells and quartz media. The
results for sand zeta potential are provided in Table 2.
2.5. Column experiments
An adjustable length chromatographic column (Omnifit,
Boonton, NJ) with a 1.5 cm inner diameter was used for the
transport experiments. The column length was adjusted to 10
cm for all experiments. More than 10 pore volumes (PV) of DI
water was pumped through the column using a syringe pump
(KD Scientific Inc., New Hope, PA), followed by at least 10 PV of
the electrolyte at the selected IS to equilibrate the column. A 4
Table 2 – Electrokinetic properties of quartz sand and E. coli O15from DLVO interaction calculation.a
Ionicstrength(mM)
Quartz surfacepotential (mV)
Outer surfacepotentialb of
E. coli O157:H7 (mV)
Ef
Untreated Proteinase KTreated
U
0.01 �68.2 �27.7 �73.1
0.1 �66.7 �3.73 �23.0
1 �51.6c �0.375 �2.86
100 �13.1c �0.004 �0.029
a The interaction energies were determined from the sum of retarded
energies. Sphere–plate and sphere–sphere models were used to calculate
values of 6.5 � 10�21 and 7.5 � 10�21 J were chosen as Hamaker constant
(A131), respectively.
b Value converted from experimentally measured electrophoretic mobili
1995).
c Values from Kim et al. (2009a).
d No Energy Barrier.
PV pulse of the cell suspension with a concentration of ca.
5 � 108 cells/mL was injected, followed by a bacteria-free
electrolyte solution (>20 PV). If needed, flow interruption and
low IS solution (i.e., DI water) flush tests were conducted when
the effluent cell concentration was close to baseline (ca. zero).
The Darcy velocity of the system was maintained at 0.1 cm/min,
and the corresponding Reynolds number and Peclet
number values were 9.8 � 10�3 and 5.3 � 10�3, respec-
tively. The effluent concentration of E. coli was measured
at 280 nm by using a UV/Vis spectrophotometer (SP-890,
Barnstead International). The column experiments were
conducted at room temperature (22–25 �C).
The profile of retained E. coli O157:H7 cells in the column
was determined after recovery of the breakthrough curve (BTC)
using the procedure of Bradford et al. (2006b). The quartz sand
was carefully excavated in 1 cm increments and placed into 50
mL tubes (Fisher Scientific) containing approximately 4 mL of
the same electrolyte solution used for column experiments.
The tubes were gently shaken for approximately 3 min by
hand, and the supernatant collected and analyzed to deter-
mine the intensity of cell fluorescence using a Turner Quan-
tech Fluorometer (FM109545, Barnstead International,
Dubuque, IA). In this case, the fluorometer was used instead of
a UV/Vis spectrophotometer to reduce background interfer-
ence that originates from sand when using the spectropho-
tometer. With the fluorometer, the impact of this background
interference was always less than 1%. The fluorometer has
a fluorescent filter set with excitation and emission wave-
lengths of 490 and 515 nm to match the fluorescence of the E.
coli O157:H7/pGFP strain. The concentration of the supernatant
was determined from the sample fluorescence value via
a concentration vs. fluorescence calibration curve. A new
calibration curve was established for every experiment. The
electrolyte volume and quartz mass in each tube was deter-
mined from a mass balance (samples were weighed before and
after oven drying the sample overnight at 130 �C). In this work
7:H7 cells as well as the height of energy barrier determined
nergy barrier height (kT )or cell–quartz interaction
Energy barrier height (kT )for cell–cell interaction
ntreated ProteinaseK Treated
Untreated ProteinaseK Treated
512.9 2423.3 363.4 2609.7
9.8 338.9 3.7 227.0
NBd 5.0 NB NB
NB NB NB NB
van der Waals interaction and electrical double layer interaction
the cell–quartz and cell–cell interaction energies, respectively. The
s for bacterium–water–quartz (A132) and bacterium–water–bacterium
ty in Table 1 based on Ohshima’s soft particle theory (Ohshima, 1994,
w a t e r r e s e a r c h 4 4 ( 2 0 1 0 ) 1 0 8 2 – 1 0 9 31086
the total mass balance is given as 100 ¼ Mrec þ Mirrev, where
Mrec and Mirrev denote the percentage of injected cells that are
recovered and irreversibly retained in the column, respec-
tively. The value of Mrec ¼MeffþMDIþMFIþMsand, where Meff,
MDI, and MFI denote the percentage of injected cells that are
recovered by integration of the BTC during the initial colloid
transport phase, the DI water flush, and the flow interruption,
respectively, and Msand is the percentage obtained from
column dissection. Hence, the percentage of injected cells that
are retained in the column is given as MDI þ MFI þ Msand þMirrev, the reversibly retained cells is given as MDIþMFIþMsand,
and the irreversibly retained cells is given as Mirrev¼ 100�Mrec.
The validity of this approach was confirmed by the relatively
good mass balance results (Mrec equaled 100.3–109.5%)
obtained under highly unfavorable conditions (IS ¼ 0.01 mM
for untreated cells, and IS of 0.01 and 0.1 mM for treated cells).
2.6. Calculation of cell–cell and cell–quartz interactionenergy profiles
DLVO theory was used to calculate the total cell–cell and cell–
quartz interaction energy as a function of separation distance.
Cell–cell interaction energy was determined by considering
the cell–cell interaction with the sphere–sphere model,
whereas cell–quartz interaction energy was calculated by
considering the system using the sphere–plate model. Due to
the presence of extracellular macromolecules on the cells, the
EPM data (Table 1) was converted to the outer surface poten-
tial using Ohshima’s soft particle theory (Ohshima, 1994,
1995). These outer surface potentials and measured zeta
potentials for the sand surfaces (Table 2) were used to calcu-
late cell–cell and cell–quartz interaction energy profiles based
on DLVO theory.
The retarded van der Waals attractive interaction energies
for sphere–sphere (FVDW-SS) and sphere–plate (FVDW-SP)
systems were calculated using Equations (1) and (2), respec-
tively (Gregory, 1981).
FVDW-SS ¼ �A131aP1aP2
6hðaP1 þ aP2Þ
�1� 5:32h
lln�
1þ l
5:32h
���1
(1)
FVDW�SP ¼ ��
A132aP
6h
��1þ
�14hl
���1
(2)
Here aP1 and aP2 are the radii of two interacting cells, and h is
the separation distance, aP is the radius of the cell interacting
with quartz surface, l is the characteristic wavelength (usually
taken as 100 nm), A is the Hamaker constant. The subscripts 1,
2, and 3 on A denote bacterium, quartz, and water, respec-
tively. The A132 value of 6.5 � 10�21 J was adopted from Red-
man et al. (2004). The value of A131 was estimated to be 7.5 �10�20 J based on the geometrical mean assumption for
Hamaker constants (van Oss, 1994), and assuming 8.4 � 10�20
and 4.6 � 10�20 J for A22 and A33 components, respectively
(Israelachvili, 1992; Visser, 1972).
The electrical double layer interaction energies for sphere–
sphere (FEDL-SS) and sphere–plate (FEDL-SP) systems were
calculated using Equations (3) and (4), respectively (Hogg et al.,
1966).
FEDL�SS ¼2paP1aP2nNkTðaP1þaP2Þk2
�42
P1þ42P1
� 24P14P2
42P1þ42
P2
ln
�1þexpð�khÞ1�expð�khÞ
�
þln½1�expð�2khÞ��
(3)
FEDL�SP ¼ p303raP
�2jPjCln
�1þ expð�khÞ1� expð�khÞ
�þ�j2
P þ j2C
ln½1� expð�2khÞ��
(4)
Here nN denotes the bulk number density of ions (N m�3,
where N is the number), k is the Boltzmann constant (J K�1), T
the absolute temperature of the system (K ), 4P1 and 4P2 are the
reduced potentials (4 ¼ zej/kT ) of two interacting cells
(dimensionless), jP and jC are the electric potentials of the cell
and the quartz (V), z is the ion valence (dimensionless), e the
electron charge (C ), 30 the permittivity of vacuum (C V�1 m�1),
3r the relative permittivity of water (dimensionless), and k
represents the Debye–Huckel reciprocal length (m�1). The
value of k can be determined by
k ¼
ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi�e2P
ni0z2i
303rkT
�s(5)
Here ni0 and zi represent number concentration and valence
of ion i in bulk solution, respectively.
3. Results and discussion
3.1. Electrokinetic property of untreated and proteinaseK treated E. coli O157:H7
The EPM for untreated cells ranged from�1.23 to�0.11 (mm/sec)/
(V/cm) when the IS ranged from 0.01 to 100 mM KCl and the pH
was 5.8. In contrast, the EPM of the proteinase K treated cells
ranged from �3.09 to �0.05 (mm/sec)/(V/cm) over these same
solution chemistry conditions (Table 1). This observation indi-
cates that the extracellular macromolecules may shift the posi-
tion of the shear plane away from the cell surface and diminish
the surface potential of E. coli O157:H7 cells as suggested by
Elimelech et al. (1995). It should be mentioned that differences in
the EPM of untreated and proteinase K treated cells were most
pronounced for IS � 1 mM (P < 0.05), whereas no statistical
difference (P > 0.05) was observed for IS of 100 mM.
3.2. Cell–cell and cell–quartz interaction energies
Table 2 presents results for cell–cell and cell–quartz interac-
tion energy calculations for both untreated and proteinase K
treated cells. No energy barrier occurs between cells and sand
surface when the IS � 1 mM, indicating that similar transport
behavior is expected for both untreated and treated cells
under these conditions. Hence, our subsequent studies dis-
cussed below for the proteinase K treated cells were only
conducted at IS of 0.1 and 0.01 mM.
Fig. 1A and B show the cell–quartz and cell–cell DLVO
interaction energy profiles, respectively, for untreated cells.
Increasing the IS compressed the diffuse double layer outside
0 50 100 150 200 250 300
Separation Distance (nm)
0
1000
2000
3000
Inte
ract
ion
Ene
rgy
(kT
)
Inte
ract
ion
Ene
rgy
(kT
)
Inte
ract
ion
Ene
rgy
(kT
)
Inte
ract
ion
Ene
rgy
(kT
)
0.01 mM0.1 mM1 mM100 mM
untreated O157:H7cell-quartz
0 50 100 150 200 250 300
Separation Distance (nm)
0
1000
2000
3000
0.01 mM0.1 mM1 mM100 mM
untreated O157:H7cell-cell
0 50 100 150 200 250 300
Separation Distance (nm)
0
1000
2000
3000
0.01 mM0.1 mM
treated O157:H7cell-quartz
0 100 150 200 250 300
Separation Distance (nm)
0
1000
2000
3000
0.01 mM0.1 mM
treated O157:H7cell-cell
50
C D
BA
Fig. 1 – Cell–quartz and cell–cell interaction energy profiles for untreated (A and B) and proteinase K treated (C and D) E. coli
O157:H7 cells as a function of separation distance at different solution IS. Surface potentials of E. coli O157:H7 and quartz
given in Table 2 were used for the calculation. Bacterial cell sizes used in calculations are adapted from Kim et al. (2009b).
w a t e r r e s e a r c h 4 4 ( 2 0 1 0 ) 1 0 8 2 – 1 0 9 3 1087
the surface of the cells and the quartz sand, and subsequently
caused a decrease in the electrostatic repulsive force between
the two surfaces. No energy barrier exists for cell–quartz and
cell–cell interactions for IS � 1 mM, indicating that the inter-
acting conditions are favorable for cell attachment to quartz
and to other cells. On the other hand, the interaction energy
profiles show an energy barrier to cell attachment to quartz and
to other cells at IS ¼ 0.1 mM (9.8 and 3.7 kT for cell–quartz and
cell–cell interactions, respectively) and IS¼ 0.01 mM (512.9 and
363.4 kT for cell–quartz and cell–cell interactions, respectively),
and no secondary energy minimum. These observations indi-
cate highly unfavorable conditions for E. coli O157:H7 deposition
to the quartz and for cell–cell interactions when the IS� 0.1 mM.
Fig. 1C and D show the cell–quartz and cell–cell DLVO
interaction energy profiles, respectively, for the proteinase K
treated cells for IS of 0.1 and 0.01 mM. Similarly to the
untreated cells, the interaction energy profiles indicate unfa-
vorable conditions for cell–quartz and cell–cell interactions at
these IS. However, much larger energy barriers are predicted
for proteinase K treated cells than untreated cells. Specifically,
an enormous energy barrier exists for cell–quartz and cell–cell
interactions at IS¼ 0.1 mM for the treated cells (338.9 and 227.0
kT for cell–quartz and cell–cell interactions, respectively) as
compared to the untreated cells (9.8 and 3.7 kT for cell–quartz
and cell–cell interactions, respectively). At an IS of 0.01 mM the
energy barrier to cell–quartz and cell–cell interactions was very
unfavorable for both untreated and treated cells (>363 kT ).
3.3. Cell aggregation
The results of aggregation experiments (average of three
replicates) with untreated cells are presented in Fig. 2.
Increasing the IS led to faster cell aggregation as suggested by
a steeper slope. This is due to cell–cell aggregates forming and
settling out of suspension. The average first-order aggregation
rate constants (k) obtained from the data for the untreated
cells were determined to be 8.8 � 10�3, 1.3 � 10�2, 2.3 � 10�2,
5.6� 10�2 h�1 at IS of 0.01, 0.1, 1, and 100 mM KCl, respectively
(Table 3). The cell aggregation rate at IS of 0.1, 1, and 100 mM is
approximately 1.5, 2.6, and 6.4 times greater than that at IS ¼0.01 mM. The fact that cell aggregation rates increase with IS
supports the possibility of cell–cell interactions, and is quali-
tatively consistent with the predicted trend of cell–cell inter-
actions shown in Fig. 1B.
Average cell aggregation results for proteinase K treated
cells at IS of 0.1 and 0.01 mM are also shown in Fig. 2. Similar
with the result for the untreated cells, increasing the IS led to
slightly faster cell aggregation rates (Table 3); however, the
0 0.5 1 1.5 2Time Elapsed (hrs)
0.90
0.92
0.94
0.96
0.98
1.00C
/C0
blank0.01 mM, untreated0.1 mM, untreated1 mM, untreated100 mM, untreated0.01 mM, treated0.1 mM, treated
Fig. 2 – Relative concentration vs. time for untreated (open)
and proteinase K treated (solid) E. coli O157:H7 cells.
Experiments were carried out at pH 5.8 and at room
temperature (22–25 8C). Error bars indicate one standard
deviation.
w a t e r r e s e a r c h 4 4 ( 2 0 1 0 ) 1 0 8 2 – 1 0 9 31088
difference was not statistically significant (P > 0.05). The
aggregation rate for the treated cells (3.3 � 10�3 and 4.8 � 10�3
h�1 at IS of 0.01 and 0.1 mM, respectively) is lower than the
untreated cells (8.8 � 10�3 and 1.3 � 10�2 h�1 at IS of 0.01 and
0.1 mM, respectively) at the same IS conditions. This obser-
vation suggests that macromolecules on the untreated cells
may still be weakly interacting even at an IS of 0.1 and 0.01
mM and enhancing cell–cell interaction.
3.4. Transport and retention behavior of untreatedE. coli O157:H7
Fig. 3 shows BTCs (Fig. 3A) and retention profiles (Fig. 3B) for
untreated E. coli O157:H7 cells at IS of 0.01, 0.1, 1, and 100 mM
Table 3 – Average first-order aggregation rate constants(k) and coefficients of linear regression values (r2) foruntreated and proteinase K treated E. coli O157:H7 cells asa function of IS.a
Ionicstrength(mM)
Untreated Proteinase K treated
k (h�1) r2 k (h�1) r2
0.01 8.8 � 10�3 �1.0 � 10�3
0.985 3.3 � 10�3 �7.7 � 10�4
0.833
0.1 1.3 � 10�2 �1.8 � 10�3
0.967 4.8 � 10�3 �1.4 � 10�3
0.918
1 2.3 � 10�2 �2.3 � 10�3
0.983 ND ND
100 5.6 � 10�2 �2.1 � 10�3
0.974 ND ND
a The values were determined by fitting first-order kinetic equa-
tion to the experimentally measured data in Fig. 2. ND Not
determined.
KCl. Mass balance information was determined from these
BTCs and the cell retention profiles, and the results are
summarized in Table 4. The quantity of cells recovered in the
column effluent (Meff) tended to decrease with increasing IS
(84.7, 9.3, 0.9, and 1.9% of total injected mass at IS of 0.01, 0.1, 1,
and 100 mM KCl, respectively), and the concentration of
retained cells correspondingly increases with IS. Differences
in the cell retention profiles between IS of 1 and 100 mM
systems shown in Fig. 3B is due to the loss of the cells eluted
by two additional steps (i.e., flow interruption and DI water
flush) after recovery of the BTC in the IS ¼ 1 mM system. Note
that no flow interruption and DI water flush tests were carried
out for the IS ¼ 100 mM system. Overall, the transport and
retention behavior is conceptually consistent with the DLVO
profile presented in Fig. 1, in that increasing the IS allowed
more cells to be retained in the column (Fig. 3A and Table 4) as
observed from other studies (Elimelech et al., 1995; Redman
et al., 2004). A close examination of this data, however, reveals
some discrepancies with DLVO predictions, which are dis-
cussed below.
The DLVO interaction energy profiles presented in Fig. 1A
and B predict that cell–quartz and cell–cell interactions are
favorable when IS � 1 mM, slightly unfavorable when IS ¼ 0.1
mM, and very unfavorable when IS ¼ 0.01 mM. Favorable
conditions imply irreversible attachment in the primary
minimum. Consistent with this prediction, almost 100% of the
injected cells were retained in the column when the IS was �1
mM (Fig. 3A). In contrast, mass balance results shown in Table
4 indicate that the percent of cells reversibly retained in the
column (Msand þ MFI þ MDI) increased with increasing IS
(conversely, 40.7, 40.7, and 45.1, and 0% of the injected cells
were irreversibly retained in the systems with IS of 100, 1, 0.1,
and 0.01 mM, respectively). It therefore appears that similar
amounts of irreversibly retained cells occurred when IS � 0.1
mM. Under favorable attachment conditions (IS of 1 and 100
mM shown in Fig. 1) irreversible attachment in the primary
minimum is expected due to the absence of an energy barrier.
Under slightly unfavorable attachment conditions (IS ¼ 0.1
mM) it is possible for some of the cells to overcome the energy
barrier (9 kT ) to attachment in the primary minimum due to
diffusion (Franchi and O’Melia, 2003) or chemical heteroge-
neity (Baygents et al., 1998; Walker et al., 2005).
Table 4 reports that 15.6, 45.6, 58.4 (MsandþMFIþMDI in this
case), and 57.4% of the injected cells in the IS of 0.01, 0.1, 1, and
100 mM systems, respectively, were reversibly recovered from
the sand after completion of the column experiments. The
cells that can be recovered after DI water flush, flow inter-
ruption, or excavating the sand in the column are weakly
associated with the solid phase. Interaction of the cells on the
quartz surface by the secondary energy minimum is unlikely
based on calculations presented in Fig. 1 and Table 2. Three
factors can likely explain the reversible cell retention
behavior, namely: (a) steric stabilization; (b) cell–cell interac-
tions; and (c) funneling of weakly associated cells to the
smallest regions of the pore space that are associated with
lower hydrodynamic forces. Each of these hypotheses will be
discussed below.
Steric stabilization of extracellular macromolecules can
limit the approach distance of cells to the quartz surface
(Elimelech et al., 1995; Israelachvili, 1992) and thereby hinder
0 10 20 30Pore Volume
0
0.2
0.4
0.6
0.8
1
C/C
0
0.01 mM0.1 mM1 mM + FI + DI100 mM
DI water flush
0 0.02 0.04 0.06 0.08 0.1Nc Ntc
-1 g-1 of sand
1
0.8
0.6
0.4
0.2
0
Dim
ensi
onle
ss D
epth
0.01 mM0.1 mM1 mM + FI + DI100 mM
A B
Fig. 3 – Breakthrough curves (A) and retention profiles (B) for untreated E. coli O157:H7 at different IS. In the legend, FI and DI
denote flow interruption and DI water flush, respectively. In Fig. 3B, the deposition profile at IS [ 1 mM was obtained after
two consecutive flow interruptions (occurring at PV of 13.3 and 16.5, respectively) and the DI water flush. The X-axis
represents the normalized concentration of the cells (the number of cells in each section, Nc, divided by the total number
injected into the column, Ntc) per gram of dry sand. The breakthrough and retention profile data for IS of 1 and 100 mM were
adapted from Kim et al. (2009a).
w a t e r r e s e a r c h 4 4 ( 2 0 1 0 ) 1 0 8 2 – 1 0 9 3 1089
attachment in the primary minimum, even under favorable
attachment conditions (IS ¼ 1 and 100 mM). For example,
Kuznar and Elimelech (2006) observed in an RSPF system that
the attachment efficiency of C. parvum oocysts was much
smaller than unity under favorable attachment conditions.
These authors attributed this unusual deposition trend to
electrosteric stabilization of the macromolecules on the
oocyst surface. As mentioned in the introduction, they did not
discuss the potential reversibility of retained oocysts by steric
repulsion since the RSPF system only captures irreversible cell
attachment (i.e., primary energy minimum interaction).
Table 4 – Mass balance results for untreated andproteinase K treated E. coli O157:H7 cells obtained fromcolumn experiments.a
Ionic strength(mM)
Meff
(%)MFI
(%)MDI
(%)Msand
(%)Mrec
(%)100 � Mrec
(%)
Untreated E. coli O157:H7
0.01 84.7 – – 15.6 100.3 <0
0.1 9.3 – – 45.6 54.9 45.1
1b 0.9 0.8c 18.7 38.9 59.3 40.7
100b 1.9 – – 57.4 59.3 40.7
Proteinase K treated E. coli O157:H7
0.01 100.3 1.1d 0.4 5.2 107.0 <0
0.1 89.7 1.7d 11.8 6.3 109.5 <0
a Meff, MDI, and MFI denote the percentage of injected cells that are
recovered by integration of the BTC during the initial colloid
transport phase, the DI water flush, and the flow interruption,
respectively, and Msand is the percentage obtained from column
dissection. Mrec represents the sum of all percentages (i.e., Meff, MFI,
MDI, and Msand).
b Data from Kim et al. (2009a).
c Percent recovery of the injected cells eluted by 1st (w3.2 PV) and
2nd (w10.4 PV) flow interruption.
d Percent recovery of the injected cells eluted by flow interruption
of 5.2 PV.
However, it is plausible that the addition of repulsive steric
force to DLVO forces (i.e., sum of van der Waals attractive and
electrostatic forces) may lead to the change in the total
interaction energy and make the conditions more unfavorable
(Hiemenz and Rajagopalan, 1997). Furthermore, the possibility
of weak interaction between cells and surfaces (i.e., the exis-
tence of secondary energy minimum interaction) may be
dependent on the magnitude of repulsive steric interaction
(Hiemenz and Rajagopalan, 1997; Yoshioka et al., 1997). In
addition to the cell–quartz steric stabilization, cell–cell steric
stabilization could also contribute to the reversible cell
retention observed under favorable conditions (i.e., 1 and 100
mM). In comparison to direct interaction of cells in the
primary minima, steric stabilization produces weak cell–
quartz or cell–cell interactions at a larger separation distance
that is controlled by the extracellular macromolecule layer.
Hence, steric stabilization of cells in the primary minima is
expected to act in a similar way as cells associated with the
secondary minima.
Fig. 1B indicates that cell–cell interactions are favorable in
the IS of 1 and 100 mM systems, but unfavorable in the IS of 0.1
mM (3.7 kT energy barrier) and 0.01 mM (363.4 kT energy
barrier) systems. The cell breakthrough results for the IS of 0.1
and 0.01 mM systems (Fig. 3A) show that the effluent
concentrations in the plateau region of the BTCs decrease
with time (i.e., ripening), indicating that cell–cell interactions
are also likely involved in the cell retention mechanism
(Bradford et al., 2007; Liu et al., 2008; Tong et al., 2008) under
these conditions. The hypothesis that cell–cell interactions
may be involved in cell retention processes is supported by
results from the aggregation experiments shown in Fig. 2 and
Table 3, as well as observations from a micromodel study (Kim
et al., 2009a) that found that E. coli O157:H7 cells were prefer-
entially attached to and coated specific sand grains under
certain solution chemistry conditions. Furthermore, the cell
retention profiles shown in Fig. 3B do not always decrease
monotonically with depth (e.g., the greatest cell retention
w a t e r r e s e a r c h 4 4 ( 2 0 1 0 ) 1 0 8 2 – 1 0 9 31090
occurred at a dimensionless depth of ca. 0.13 for the IS ¼ 100
mM system). Such non-monotonic (bell-shaped) retention
profiles are not consistent with the exponential decay in the
amount of attached cells with distance predicted by classic
filtration theory (Yao et al., 1971). These results suggest that
the cells weakly associated with quartz surface or other cells
by steric interaction may detach in the presence of hydrody-
namic drag and subsequently be retained down-gradient. It
should be noted that aggregated cells would be more vulner-
able to hydrodynamic shear than a singe cell (Torkzaban et al.,
2007). This hypothesis is supported by the results from Tong
et al. (2005) and Bradford et al. (2006b). For example, Tong et al.
(2005) observed non-monotonic retention profiles from
a column study using Comamonas DA001 under favorable
conditions. They also attributed the unusual shape of the
retention profiles to weak interaction by steric repulsion as
well as detachment of the cells due to hydrodynamic drag.
These hypotheses were supported by the observed migration
of the peak retained cell concentration down-gradient with
increasing elution time and flow rate. Bradford et al. (2006b)
also reported non-monotonic retention profiles under unfa-
vorable conditions using the same E. coli O157:H7 strain used
in this study. They attributed the non-monotonic retention
profile to aggregation, breakage, and release of the cells in low
velocity regions found near grain junctions, small pores, and
pore constrictions. Potential explanations for cell–cell inter-
actions in the IS of 0.1 and 0.01 mM systems include polymer
bridging (Jucker et al., 1998; Liu et al., 2007, 2008) and biological
cell–cell regulation (i.e., quorum sensing) (Parsek and Green-
berg, 2000). Polymer bridging may occur under low IS condi-
tions when the polymeric layers extend into the bulk solution
(Abu-Lail and Camesano, 2003; Liu et al., 2007, 2008) and
interact with high affinity surfaces (Liu et al., 2007). Indeed,
the E. coli O157:H7 cells possess extracellular polymeric
materials, and therefore polymer bridging interaction is
plausible (Kim et al., 2009b). At the present, however, the exact
mechanism for cell–cell interactions under these conditions is
uncertain.
0 10 20 30Pore Volume
0
0.2
0.4
0.6
0.8
1
C/C
0
0.01 mM + FI + DI0.1 mM + FI + DI
DI water flushfor 0.01 mM
DI water flush for 0.1 mM
A B
Fig. 4 – Breakthrough curves (A) and retention profiles (B) for pro
legend, FI and DI denote flow interruption and DI water flush,
obtained after a flow interruption and DI water flush. The X-ax
number of cells in each section, Nc, divided by the total numbe
Many researchers have demonstrated that low velocity
regions that are induced by the pore structure (e.g., grain
junctions, small pore spaces, and pore constrictions) can
enhance colloid/cell retention in porous media (Bradford et al.,
2006b, 2007; Tong et al., 2008; Torkzaban et al., 2008) under
unfavorable attachment conditions. In the present study, low
velocity regions are also likely to be playing an important role
in the retention of the E. coli O157:H7 cells. The result for IS¼ 1
mM, for instance, shows that a significant amount of the
injected cells (Msand ¼ 38.9%) were still recovered from the
column after conducting flow interruption and flushing the
system with DI water. Recent literature has demonstrated that
weakly associated cells and colloids in the secondary
minimum may be funneled by hydrodynamic forces to
regions that are associated with lower hydrodynamic forces
and therefore enhanced retention (Bradford et al., 2007; Tong
et al., 2008). Similarly, we hypothesize herein that cells weakly
associated with the solid phase by steric stabilization may
also be funneled to and retained in low velocity regions.
3.5. Transport and retention behavior of proteinase Ktreated E. coli O157:H7
Fig. 4 shows BTCs (Fig. 4A) and retention profiles (Fig. 4B) for
proteinase K treated E. coli O157:H7 cells when the IS was 0.01
and 0.1 mM. Similar to the BTC results for untreated E. coli
O157:H7 cells (Fig. 3A), more cells were retained in the column
as IS increases (i.e., cell breakthrough of 100.3 and 89.7% of the
injected cells in IS of 0.01 and 0.1 mM systems, respectively).
A significantly smaller amount of the proteinase K treated
cells were retained in the column as compared to the
untreated ones (Table 4), indicating that the extracellular
macromolecules of E. coli O157:H7 cells enhance the cell
retention. The notably different cell retention between the
proteinase K treated and untreated cells (IS¼ 0.1 mM) is due to
electrostatic interactions of the untreated (9.8 kT ) being much
less than the proteinase K treated (338.9 kT ) cells (Fig. 1 and
Table 2). Cell–cell interactions were also diminished by the
0 0.02 0.04 0.06 0.08 0.1Nc Ntc
-1 g-1 of sand
1
0.8
0.6
0.4
0.2
0
Dim
ensi
onle
ss D
epth
0.01 mM + FI + DI0.1 mM + FI + DI
teinase K treated E. coli O157:H7 cells at different IS. In the
respectively. The cell deposition profiles in Fig. 4B were
is represents the normalized concentration of the cells (the
r injected into the column, Ntc) per gram of dry sand.
w a t e r r e s e a r c h 4 4 ( 2 0 1 0 ) 1 0 8 2 – 1 0 9 3 1091
removal of the extracellular macromolecules on the E. coli
O157:H7 cells. This conclusion is supported by the difference
in the DLVO interaction energies (Fig. 1B and D), the cell
aggregation results (Fig. 2 and Table 3), and differences in the
shape of the BTC for untreated (Fig. 3A) and proteinase K
treated cells (Fig. 4A). In particular, notice that the BTC for
proteinase K treated cells (Fig. 4A) does not exhibit ripening
that occurred for the untreated cells (Fig. 3A).
The cell–quartz interaction energy shown in Fig. 1C indi-
cates that conditions are highly unfavorable for irreversible
cell attachment in the primary minimum (height of energy
barrier ¼ 2423.3 and 338.9 kT at 0.01 and 0.1 mM, respectively)
and no secondary energy minimum was present. However,
mass balance results presented in Table 4 show that reversible
retention of the treated cells occurred in the 0.01 and 0.1 mM
systems. Plausible mechanisms for the reversible cell reten-
tion under these conditions include interaction by polymer
bridging (Jucker et al., 1998; Liu et al., 2007, 2008) and retention
induced by the pore structure and/or the quartz surface
roughness (Bradford et al., 2006a,b, 2007; Tufenkji and
Elimelech, 2004; Tufenkji et al., 2004). Mass balance results
show that approximately 11.8% of the injected cells (MDI) were
eluted by switching the electrolyte to DI water in the IS ¼ 0.1
mM system, indicating the presence of a weak cell–quartz
interaction. Here, it should be noted that the treated cells may
still possess polysaccharide chains, which are not linked with
protein structure and may cause cell retention to some extent,
since proteinase K cleaves the macromolecules associated
with protein structure (Kuznar and Elimelech, 2006; Kim et al.,
2009b). The fact that the reversible cell retention is likely due to
the polymer bridging interaction can also be supported by the
tailing behavior observed in the BTC for the IS ¼ 0.1 mM
system. In particular, notice that only 1.7% (MFI) of the injected
cells were released by flow interruption, whereas 11.8% (MDI)
were released by flushing with DI water. If the tailing was
controlled by cell diffusion, then the flow interruption test
results would have shown larger amounts of cell release. The
results indicate that the tailing is likely associated with
the detachment of the cells due to the hydrodynamic shear force
(Li et al., 2005; Bradford et al., 2007; Liu and Li, 2008), which may
otherwise be weakly attached to the quartz surface by polymer
bridging. It should be noted that even after DI water flushing,
approximately 5.2 and 6.3% of the injected cells (Msand) were
reversibly recovered from the quartz sand at IS of 0.01 and 0.1
mM, respectively. This retention likely occurred as a result of the
pore structure and/or surface roughness, which may serve as
hydrodynamically favorable regions for cell retention (Vaidya-
nathan and Tien, 1988; Torkzaban et al., 2008).
4. Conclusions and implications
Comparison of the transport and retention behavior between
macromolecule rich and deficient cells indicate that patho-
genic E. coli O157:H7 retention in saturated porous media can
be significantly influenced by macromolecule-mediated
interactions. It was found that the macromolecule-related
interactions are strongly linked to solution IS. Classic DLVO
theory, which does not account for macromolecule-mediated
interactions such as polymer bridging or steric repulsion, was
found to be insufficient to accurately predict the retention of E.
coli O157:H7 cells in porous media. Evidence to this effect was
the fact that a significant amount of cells were retained under
unfavorable conditions with no secondary energy minimum,
and irreversible cell retention was hindered under favorable
conditions.
Significantly enhanced retention with time (i.e., ripening)
and a faster aggregation rate for untreated cells suggests that
cell–cell and cell–quartz interactions are also important
mechanisms for the retention of E. coli O157:H7 cells in porous
media. Furthermore, the non-monotonic cell retention profile
under favorable conditions, which deviates from classic
filtration theory, indicates that cells may be weakly associated
with other cells or quartz surface due to steric interactions
and susceptible to removal by hydrodynamic forces. Both of
these processes are commonly neglected in classical theories
of cell interaction and retention. Caution is therefore war-
ranted in predicting the travel distance of pathogens in
(drinking or riverbank) water filtration facilities using current
theoretical models. For example, neglecting cell–cell interac-
tions and steric interactions in travel distance predictions
may lead to an overestimation and underestimation, respec-
tively. Furthermore, biofilm formation developed by cell–cell
interactions might cause significant costs in long-term oper-
ation of treatment systems.
Acknowledgements
This research was funded by the USDA CSREES NRI (Grant #:
2006-02541). Permission from the American Chemical Society
to use the electrophoretic mobility data (Kim et al., 2009b) that
is presented in Table 1 is gratefully acknowledged. The
breakthrough and retention profile data for IS of 1 and 100 mM
systems (Kim et al., 2009a) shown in Fig. 3 is used with kind
permission of Springer Science and Business Media.
r e f e r e n c e s
Abu-Lail, N.I., Camesano, T.A., 2003. Role of ionic strength on therelationship of biopolymer conformation, DLVO contributions,and steric interactions to bioadhesion of Pseudomonas putidaKT2442. Biomacromolecules 4, 1000–1012.
Baygents, J.C., Glynn Jr., J.R., Albinger, O., Biesemeyer, B.K.,Ogden, K.L., Arnold, R.G., 1998. Variation of surface chargedensity in monoclonal bacterial populations: implications fortransport through porous media. Environ. Sci. Technol. 32,1596–1603.
Boulos, L., Prevost, M., Barbeau, B., Coallier, J., Desjardins, R., 1999.LIVE/DEAD BacLight: application of a new rapid stainingmethod for direct enumeration of viable and total bacteria indrinking water. J. Microbiol. Meth. 37, 77–86.
Boyce, T.G., Swerdlow, D.L., Griffin, P.M., 1995. Current concepts:Escherichia coli O157:H7 and the hemolytic-uremic syndrome.N. Engl. J. Med. 333, 364–368.
Bradford, S.A., Simunek, J., Battahar, M., van Genuchten, M.T.,Yates, S.R., 2006a. Significance of straining in colloiddeposition: evidence and implications. Water Resour. Res. 42,W12S15.
w a t e r r e s e a r c h 4 4 ( 2 0 1 0 ) 1 0 8 2 – 1 0 9 31092
Bradford, S.A., Simunek, J., Walker, S.L., 2006b. Transport andstraining of E. coli O157:H in saturated porous media. WaterResour. Res. 42, W12S12. doi:10.1029/2005WR004805.
Bradford, S.A., Torkzaban, S., Walker, S.L., 2007. Coupling of physicaland chemical mechanisms of colloid straining in saturatedporous media. Water Res. 41, 3012–3024.
Chen, G., Walker, S.L., 2007. The role of solution chemistry andion valence on the adhesion kinetics of groundwater andmarine bacteria. Langmuir 2 (3), 7162–7169.
Dean-Nystrom, E.A., Melton-Celsa, A.R., Pohlenz, J.F.L., Moon, H.W.,O’Brien, A.D., 2003. Comparative pathogenicity of Escherichia coliO157 and intimin-negative non-O157 shiga toxin-producing E.coli strains in neonatal pigs. Infect. Immun. 71, 6526–6533.
Derjaguin, B.V., Landau, L.D., 1941. Theory of stability of highlycharged lyophobic sols and adhesion of highly chargedparticles in solutions of electrolytes. Acta Physicochim. USSR14, 633–662.
Elimelech, M., Gregory, J., Jia, X., Williams, R.A., 1995. ParticleDeposition and Aggregation: Measurement, Modeling andSimulation. Butterworth-Heinemann, Oxford, England.
Frank, J.F., 2001. Microbial attachment to food and food contactsurfaces. Adv. Food Nutr. Res. 43, 320–370.
Franchi, A., O’Melia, C.R., 2003. Effects of natural organic matterand solution chemistry on the deposition and reentrainmentof colloids in porous media. Environ. Sci. Technol. 37,1122–1129.
Gargiulo, G., Bradford, S.A., Simunek, J., Ustohal, P., Vereecken, H.,Klumpp, E., 2007. Bacteria transport and deposition underunsaturated conditions: the role of the matrix grain size andthe bacteria surface protein. J. Contam. Hydrol. 92, 255–273.
Ginn, T.R., Wood, B.D., Nelson, K.E., Schiebe, T.D., Murphy, E.M.,Clement, T.P., 2002. Processes in microbial transport in thenatural subsurface. Adv. Water Res. 25, 1017–1042.
Gregory, J., 1981. Approximate expressions for retarded van derWaals interaction. J. Colloid Interface Sci. 83, 138–145.
Hiemenz, P.C., Rajagopalan, R., 1997. Principles of Colloid andSurface Chemistry, third ed. Marcel Dekker, New York.
Hogg, R.I., Healey, T.W., Fuerstenau, D.W., 1966. Mutual coagulationof colloidal dispersions. Trans. Faraday Soc. 62, 1638–1651.
Israelachvili, J.N., 1992. Intermolecular and Surface Forces.Academic Press, Amsterdam, Boston.
Jin, Y., Flury, M., 2002. Fate and transport of viruses in porousmedia. Adv. Agron. 77, 39–102.
Jucker, B.A., Zehnder, A.J.B., Harms, H., 1998. Quantification ofpolymer interactions in bacterial adhesion. Environ. Sci.Technol. 32, 2909–2915.
Kaper, J.B., Karmali, M.A., 2008. The continuing evolution ofa bacterial pathogen. Proc. Natl. Acad. Sci. 105 (12), 4535–4536.
Kim, H.N., Bradford, S.A., Walker, S.L., 2009a. Escherichia coli O157:H7 transport in saturated porous media: role of solutionchemistry and surface macromolecules. Environ. Sci. Technol.43, 4340–4347.
Kim, H.N., Hong, Y., Lee, I., Bradford, S.A., Walker, S.L., 2009b.Surface characteristics and adhesion behavior of Escherichiacoli O157:H7: role of extracellular macromolecules.Biomacromolecules 10, 2556–2564.
Kuznar, Z.A., Elimelech, M., 2005. Role of surface proteins in thedeposition kinetics of Cryptosporidium parvum oocysts.Langmuir 21, 710–716.
Kuznar, Z.A., Elimelech, M., 2006. Cryptosporidium oocyst surfacemacromolecules significantly hinder oocyst attachment.Environ. Sci. Technol. 40, 1837–1842.
Kuznar, Z.A., Elimelech, M., 2007. Direct microscopic observationof particle deposition in porous media: role of the secondaryenergy minimum. Colloid Surf. Physicochem. Eng. Aspect. 294,156–162.
Law, D.J., 2000. Virulence factors of Escherichia coli O157 and otherShiga toxin-producing E. coli. J. Appl. Microbiol. 88, 729–745.
Li, X., Zhang, P., Lin, C.L., Johnson, W.P., 2005. Role ofhydrodynamic drag on microsphere deposition and re-entrainment in porous media under unfavorable conditions.Environ. Sci. Technol. 39, 4012–4020.
Litton, G.M., Olson, T.M., 1993. Colloid deposition rates on silicabed media and artifacts related to collector surfacepreparation methods. Environ. Sci. Technol. 27, 185–193.
Liu, Y., Yang, C.H., Li, J., 2007. Influence of extracellular polymericsubstances on Pseudomonas aeruginosa transport and depositionprofiles in porous media. Environ. Sci. Technol 41, 198–205.
Liu, Y., Li, J., 2008. Role of Pseudomonas aeruginosa biofilm in theinitial adhesion, growth and detachment of Escherichia coli inporous media. Environ. Sci. Technol. 42, 443–449.
Liu, Y., Yang, C.-H., Li, J., 2008. Adhesion and retention ofa bacterial phytopathogen Erwinia chrysanthemi in biofilm-coated porous media. Environ. Sci. Technol. 42, 159–165.
Mead, P.S., Slutsker, L., Dietz, V., McCaig, L.F., Bresee, J.S.,Shapiro, C., Griffin, P.M., Tauxe, R.V., 1999. Food-related illnessand death in the United States. Emerg. Infect. Dis. 5, 607–625.
Oh, Y.J., Jo, W., Yang, Y., Park, S., 2007. Influence of cultureconditions on Escherichia coli O157:H7 biofilm formation byatomic force microscopy. Ultramicroscopy 107, 869–874.
Ohshima, J., 1994. Electrophoretic mobility of soft particles.J. Colloid Interface Sci. 163, 474–483.
Ohshima, H., 1995. Electrophoretic mobility of soft particles.Colloid Surf. Physicochem. Eng. Aspect. 103, 249–255.
Parsek, M.R., Greenberg, E.P., 2000. Acyl-homoserin lactonequorum sensing in Gram-negative bacteria: a signalingmechanism involved in association with higher organisms.Proc. Natl. Acad. Sci. 97, 8789–8793.
Pembrey, R.S., Marshall, K.C., Schneider, R.P., 1999. Cell surfaceanalysis techniques: what do cell preparation protocols do tocell surface properties? Appl. Environ. Microbiol. 65, 2877–2894.
Redman, J.A., Walker, S.L., Elimelech, M., 2004. Bacterial adhesionand transport in porous media: role of the secondary energyminimum. Environ. Sci. Technol. 38, 1777–1785.
Rijnaarts, H.H.M., Norde, W., Lyklema, J., Zehnder, A.J.B., 1999. DLVOand steric contributions to bacterial deposition in media ofdifferent ionic strengths. Colloid Surf.B: Biointerfaces 14,179–195.
Ryu, J.-H., Beuchat, L.R., 2005. Biofilm formation by Escherichia coliO157:H7 on stainless steel: effect of exopolysaccharide andcurli production on its resistance to chlorine. Appl. Environ.Microbiol. 71, 247–254.
Schijven, J.K., Hassanizadeh, S.M., 2000. Removal of viruses bysoil passage: overview of modeling, processes, andparameters. Crit. Rev. Environ. Sci. Technol. 30, 49–127.
Seltmann, G., Holst, O., 2002. The Bacterial Cell Wall. Springer,Berlin, New York.
Shen, C., Li, B., Huang, Y., Jin, Y., 2007. Kinetics of coupledprimary- and secondary-minimum deposition of colloidsunder unfavorable chemical conditions. Environ. Sci. Technol.41, 6976–6982.
Tong, M., Li, X., Brow, C.N., Johnson, W.P., 2005. Detachment-influenced transport of an adhesion-deficient bacterial strainwithin water-reactive porous media. Environ. Sci. Technol. 39,2500–2508.
Tong, M., Ma, H., Johnson, W.P., 2008. Funneling of flow intograin-to-grain contacts drives colloid-colloid aggregation inthe presence of an energy barrier. Environ. Sci. Technol. 42,2826–2832.
Torkzaban, S., Bradford, S.A., Walker, S.L., 2007. Resolving thecoupled effects of hydrodynamics and DLVO forces on colloidattachment in porous media. Langmuir 23, 9652–9660.
Torkzaban, S., Tazehkand, S.S., Walker, S.L., Bradford, S.A., 2008.Transport and fate of bacteria in porous media: coupledeffects of chemical conditions and pore space geometry.Water Resour. Res. 44, W04403. doi:04410.01029/02007WR006541.
w a t e r r e s e a r c h 4 4 ( 2 0 1 0 ) 1 0 8 2 – 1 0 9 3 1093
Tufenkji, N., Elimelech, M., 2004. Deviation from classical colloidfiltration theory in the presence of repulsive DLVOinteractions. Langmuir 20, 10818–10828.
Tufenkji, N., Miller, G.F., Ryan, J.N., Harvey, R.W., Elimelech, M.,2004. Transport of Cryptosporidium oocysts in porous media:role of straining and physicochemical filtration. Environ. Sci.Technol. 38, 5932–5938.
Vaidyanathan, R., Tien, C., 1988. Hydrosol deposition in granularbeds. Chem. Eng. Sci. 43, 289–302.
van Oss, C.J., 1994. Interfacial Forces in Aqueous Media. MarcelDekker, New York.
Verwey, E.J.W., Overbeek, J.T.G., 1948. Theory of the Stability ofLyophobic Colloids. Elsevier, Amsterdam.
Visser, J., 1972. On Hamaker constants: a comparison betweenHamaker constants and Lifshitz–van der Waals constants.Adv. Colloid Interface Sci. 3, 331–363.
Voloshin, S.A., Kaprelyants, A.S., 2004. Cell–cell interactionsin bacterial populations. Biochemistry (Mosc) 69,1268–1275.
Walker, S.L., Redman, J.A., Elimelech, M., 2005. Influenceof growth phase on bacterial deposition: interactionmechanism in packed bed column and radial stagnationpoint flow systems. Environ. Sci. Technol. 39,6405–6411.
Wingender, J., Neu, T.R., Flemming, H.C., 1999. MicrobialExtracellular Polymeric Substances. Springer, Berlin.
Yao, K.M., Habibian, M.T., O’Melia, C.R., 1971. Water andwastewater treatment filtration: concepts and applications.Environ. Sci. Technol. 5, 1105–1112.
Yoshioka, K., Sakai, E., Daimon, M., Kitahara, A., 1997. Role ofsteric hindrance in the performance of superplasticizers forconcrete. J. Am. Ceram. Soc. 80, 2667–2771.